Research On Modular Algorithm For Weighted Human Brain Structure Networks

The human brain is one of the most complex systems all over the world.Researchers have been dedicated to study and explore the working principle and operating mechanism of the human brain by using new technologies.In recent years,reconstruction technology of human brain structural network based on nuclear magnetic resonance imaging is increasingly mature,and analyzing the brain network by using graph theory and complex network theory is also becoming the focus of the brain science research.Human brain structural network is a complex network.There are some module structures which play an important role in the overall operation of the brain.At present,most researches focused on the module partition method of binary brain structural network.The binary brain structural network only shows whether there is a connection relationship between brain regions in general,but the weight human brain structural network based on physiological information of human brain can express more specific relationship.It makes more sense for the module structure partition on the basis of that.It is carried out the research for the weight network modular algorithm using the weight human structural network as the object in the paper.In the first,it is finished that the construction of binary human brain structural network based on nuclear magnetic resonance imaging data and the weight human brain structural network.And march on the module partition and the analysis of the results with the Fast Newman algorithm for the binary human brain structural network.After that,it is carried out the research for module structure partition method of weight human structural network on the basis of that.And it is come up with a weight Fast Newman modular algorithm based on the thought of the condensed nodes.The algorithm constructs a weight modularity evaluation index based on the single brain node's weight and the network's total weight,and uses its increment as the measure to determine whether the brain node is merged to realize the module division.The algorithm made a comparison with the binary human brain structural network modular algorithm and the existing weight network modular algorithm.The results show that it is get the higher modularity using the method in the paper and the module structure is more consistent with the known physiological characteristics of the human brain.Finally,the method in the paper is applied to the data of the schizophrenic and the healthy patients.The contrast experiment showed that there were differences in the modular structure between the two groups of subjects.